#164164th of 218 in Coding agents and AI editors

orbital

Project agent that shares context across Claude Code, Codex and Gemini CLI

Stars
976
License
GPL-3.0
Last commit
Nov 2026
Last release
Nov 2026
Language
Python
Stars, last 29 weeks+958
22 Apr 2026: 18 stars29 Apr 2026: 21 stars6 May 2026: 24 stars13 May 2026: 24 stars20 May 2026: 24 stars27 May 2026: 24 stars3 Jun 2026: 24 stars10 Jun 2026: 24 stars17 Jun 2026: 24 stars24 Jun 2026: 34 stars1 Jul 2026: 38 stars8 Jul 2026: 79 stars15 Jul 2026: 99 stars22 Jul 2026: 103 stars29 Jul 2026: 112 stars5 Aug 2026: 132 stars12 Aug 2026: 139 stars19 Aug 2026: 230 stars26 Aug 2026: 235 stars2 Sept 2026: 253 stars9 Sept 2026: 259 stars16 Sept 2026: 266 stars23 Sept 2026: 299 stars30 Sept 2026: 309 stars7 Oct 2026: 309 stars14 Oct 2026: 333 stars21 Oct 2026: 946 stars28 Oct 2026: 965 stars4 Nov 2026: 969 stars11 Nov 2026: 976 stars

18 on 22 Apr 2026976 on 11 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Orbital is a Windows and macOS desktop app that runs a project agent over a local workspace folder. It keeps state, decisions, lessons and a task queue as plain files (PROJECT_STATE.md, DECISIONS.md, LESSONS.md, queue.json) and delegates work to Claude Code, Codex, Gemini CLI, Cursor and other CLI agents. When one agent hits a usage limit, another can continue from the same project context.

Who it is for: Developers who switch between several coding agents on one project

Strengths

  • Context lives as plain files in the project folder, readable by any agent
  • Dispatches to Claude Code (SDK), Codex (app-server), Cursor (ACP), and Gemini CLI or Aider (PTY)
  • Queued tasks must end Completed or Blocked; agents that stop without a verdict are re-prompted
  • Runs locally with your own API key; installers need no Python or Node

Weaknesses

  • GPL-3.0 license limits embedding in proprietary products
  • Installers only for Windows and macOS; no Linux build or Docker image mentioned
  • Full autonomy is on the roadmap; today it is scheduler-driven
  • Delegating needs the worker agent CLIs installed and signed in separately

What it needs

  • no GPU
  • Needs LLM provider API key, Claude Code, Codex, Gemini CLI or other agent CLIs (optional workers)
  • Models: DeepSeek, Anthropic, OpenAI, Moonshot

What changed

From the release notes. No flag means the note does not say so, not that the upgrade is safe.

Also in Coding agents and AI editors

See all 218
Also in Coding agents and AI editors
RankProjectScore
1claw-codeRust CLI coding agent harness using Anthropic and OpenAI-compatible APIs194.8k stars+724 (+0.4%) in 30 daysMIT96 out of 100
2t3codeDesktop, web and mobile front end for controlling local coding agent CLIs Live demo ↗ (opens in a new tab)57.9k stars+30,427 (+111%) in 30 daysMIT95 out of 100
3codewhaleTerminal coding agent in Rust that works with 40+ model providers41.2k stars+331 (+0.8%) in 30 daysMIT94 out of 100
#4ArchonYAML workflow engine that runs AI coding agents through fixed dev steps23.9k starsMIT93 out of 100
#5piMinimal terminal coding agent extensible with skills and packages121.6k stars+7,287 (+6.4%) in 30 daysMIT92 out of 100